Bug 49086 – Incorrect image map used when multiple maps have the same name

• Status: RESOLVED
• Resolution: FIXED
• Component: Images

[Darin Adler]

Incorrect image map used when multiple maps have the same name

[Darin Adler]

Created attachment 73104 [details]
Patch

[Darin Adler]

<rdar://problem/7627800>

[Alexey Proskuryakov]

Comment on attachment 73104 [details]
Patch

It seems a little strange that image map names are tracked in the same way as element IDs, but names in general are handled elsewhere in a very different manner (see e.g. JSHTMLDocument::nameGetter()). I guess that's fine as long as we always want the duplicate that's first in document order, but it's still surprising. Perhaps I don't understand what's so special about map names - why is there even HTMLMapElement::m_name when the same value is in an attribute, perhaps lowercased?

The test needs a subtest to make sure that the latest dynamically added map doesn't get precedence (note that subtest 5 re-adds the first element, so it doesn't test for that). It also needs an XHTML version, making sure that names are case sensitive.

Do all tests pass in Firefox?

[Darin Adler]

(In reply to comment #3)
> Do all tests pass in Firefox?

They do not. Firefox has different behavior. But HTML5 specifies the behavior the test asks for.

[Darin Adler]

(In reply to comment #3)
> Why is there even HTMLMapElement::m_name when the same value is in an attribute, perhaps lowercased?

There’s no good reason for that. There was a lot of that in the DOM implementation in the old days and I’ve removed a lot of it over time. It could and probably should be fixed.

> The test needs a subtest to make sure that the latest dynamically added map doesn't get precedence

Easy to add. Will do.

> It also needs an XHTML version, making sure that names are case sensitive.

Much less pleasant to add. Will do.

[Darin Adler]

Committed r71543: <http://trac.webkit.org/changeset/71543>